Fortepiano touchweight

[William E. Darst]

     As regards my previous post on this subject all I can say is mea
culpa!  I should have said that by placing a weight in the back of the key
the downweight is brought to or decreased to 15 grams (thereby increasing
upweight).  Thankyou Ken.

     Last week I received, out of the blue, a message from Paul Poletti, a
fortepiano maker whom I hadn't heard from in a long time.  He used to make
instruments in Modesto, Calif.  Evidently, he saw a post of mine on another
list and decided to contact me.  He is presently in Belgium building and
restoring fortepianos.  So, I put the touchweight question to him and got
the following:  (this is a bit long)

Date: 23 Feb 95 20:04:59 EST
From: Paul Poletti <100407.2266@compuserve.com>
To: Bill Darst <darst@humanitas.ucsb.edu>
Subject: Re: Fortepiano touchweight

Hi Bill,
        Interesting question. It's 1 in the morning, and I have just returned
from a typical European "pub crawl" evening, so the synapses are not firing
too clearly due to a fresh and rather generous lubrication with Belgian
beer (the best in the world, I'm afraid), but I'll try to get a coherent
answer off.
        I don't know if you know it or not, but I am considered to be a sort of
expert on Viennese actions among the builders here, mostly because I wrote
a rather massive spreadsheet which accurately predicts the functional
characteristics of the Prellzungenmechanik by geometric modeling, and then
proceeded to plug every original instrument I could my hands on into it to
see what came out. Not to blow my own horn , but just to let you know that
I've seen a lot of actions. Combined with Michael Latcham's extensive
research, which he has fed to me on a regular basis, I think I've either
examined personally or been given very good info on well over 100 extant
actions. I have *never* seen any lead in any Viennese pre-1830 action that
I could even begin to think was original. Nor have I seen any evidence of
undercutting, as is often found in harpsichords, beyond a very slight
undershaving on the sides only of natural heads, probably more for
clearance than balance. Walter does a very curious thing which gives a good
hint to the general practice: as you know, the big problem with laying out
a keyboard is that C-E is three naturals and five tails (including sharps)
and F-B is four naturals and 7 tails, and unfortunately, 3/5 does not equal
4/7. Thus the problem is how to space the sharps relative to naturals in
the two different sections. Two solutions have been found in the history of
keyboards: (1) divide the back portion of the naturals in the two sections
equally, thus your C-D-E will have slightly wider tails than F-G-A-B; (2)
keep C & E the same width as F through B and let all the error accumulate
in the D tail, which is the normal harpsichord solution. When you do this,
you end up with a noticeably wide D tail,which is also heavier. Thus on
harpsichords you
>often see more undercutting to balance the D key, since the tails at the action
>proper are all equal. But Walter keeps his D extra wide all the way back,
>even pinning the bridge and cutting the damper rail extra wide around the
>note D.  Unfortunately, the MINe109 drawing doesn't show this on the
>keyboard, though if you look very closely at the spacing of the balance
>pins and draw the strings between nut and bridge pin, it jumps right out
>at you. This is an "acid test" for a real Walter (or the accuracy of a
>copy): just look carefully at the string spacing around every D.  Why does
>he do it? Obviously, to let the wide key balance with itself without any
>undercutting.
        We have at this moment in our shop three original instruments: a
Graf 6 1/2 from about 1825, a Fritz 6 1/2 from the same period, and an
early probably Brodmann 6 octave c.1810?. None have either any lead or
undercutting. Thus it appears that the only thing that gives balance in the
Prellzungenmechanik is the weight of the key-mounted action components
themselves. This of course gives an extremely light action, and the
keyboard must be totally free and without friction in and of itself. I
personally think that slow repetition in these actions is mostly due to
either too much friction between the pawl (some call it an "escapement
lever", though it is neither an "escapement" nor a "lever" as either are
defined by classical mechanics) and the beak leather and/or pawl springs
which are too strong.
        One reads endlessly that the English actions were heavier. I don't
measure balance weight, since I just use the same materials as found in the
originals and duplicate dimensions and figure balance will take care of itself
(which is what they seem to have assumed). However, I do measure functional
weight, which is the amount of weight needed to raise the hammer through
the moment of escapement, without damper (dampers raised by the pedal or
knee lever during the test). The test weight is placed upon the key, which
is restrained, and then the key is released, which means there is no
kinetic energy of the test weight itself affecting the final results. I
have found most Prellzungenmechaniks to have functional weights of 35 to 42
grams in the bass and 21 to 27 grams in the treble, from late 18th-cent 5
octave instruments up to and including late 6 1/2 octave instruments. I
recently had the pleasure of seeing a Broadwood 5 1/2 octave from 1792
(positive date) which is in mint condition, unrestored. It had a key dip of
6.5mm consistently and a weight about 5 grams lighter than the Prell
actions. But the keys had lead in the front (!) which seemed to be
original. Thus even though the balance and functional weight was less, the
inertia of the lead made the action *feel* heavier.
        Another important point is that all, absolutely all, Viennese fortepiano
keyboards I have ever seen are made of spruce (or a similar evergreen) and
NOT linden (or basswood). A very telling instrument is the Stein 1781
claviorganum (Goteborg museum, currently on loan to Haags Gemeentemuseum),
which has two keyboards, one for the piano and one just below it for the
organ (which, by the way, slides in and out of the keywell and acts as the
sled for raising the piano action above). The organ keyboard is of  linden,
and the piano keyboard is of spruce. Here again, we see a very subtle
refinement in order to make the piano action as fleet and light as
possible. My experiments with sitka and bass while I was still in the
states indicated that sitka was about 20% lighter than bass.
        Most modern players are unable to handle real Prellzungenmechaniks
because they are so incredibly light. Thus the need for lead or
undercutting. I would say copy the action as it is. If it seems too light,
live with it for a spell. If it doesn't return well or if repitition is
bad, check your Schnabelluft: a healthy 1mm is needed. Also make sure your
key is totally frictionless, check for too strong pawl springs, polish the
pawl faces and put talc or ground soapstone on the beak leather, and most
important, make sure the player is picking his fingers ALL THE WAY UP!!!
        That's all for the moment. If I think of anything else, I'll send it
along later. I also have *extensive* studies in depth of touch, but that is
another story.

Paul

Bill Darst, RPT
Music Dept
UC Santa Barbara